Route recalculation, an imperfect science

Wapping

Moderator
Moderator
Joined
Oct 21, 2005
Messages
78,793
Reaction score
3,563
Location
Wapping, London
We are just coming to the end of a 2500 mile trip around Germany.

I created all the routes which I then shared with my two friends, both running identical Garmin latest generation GPS devices. The routes ran perfectly from day one on my Nav V; they also ran very well on my two friends' devices, once we had sorted out their preference settings. There were some differences between us, not least when it came to recalculation.

When I created the routes I asked BaseCamp to take me from hotel to hotel, A to B. I then dragged the route around so that it took me down the roads I wanted to take, using shaping points each time, not waypoints. I therefore had a final route for each day A to B with just shaping points inbetween, which I then shared with my two friends. Their two devices are set up indenticaly with each other and have a mapset one generation newer than mine.

On import, their routes were 99% the same as mine. There were occasional differences, which occurred where there two devices (using one generation younger maps) chose slightly different routes between some of the shaping points. Nothing too bad, just some slight variations. I put this down to their maps being one generation different and / or a maybe a slight difference in the software between their my Nav V and their Garmin equivalents. What was more interesting perhaps is that my two friends' devices (set up identicaly) sometimes differed in their routes between them, which suggests to me that there is sometimes a slight variation in route calculation / displaying even between otherwise supposedly identical devices.

From time to time I, who was leading each days' rides, would go 'off route', to find a coffee stop say or because of roadworks and a deviation of perhaps a couple of miles. I leave my auto-recalculation set to prompted, invariably choosing 'No' when asked whether I would like the device to recalcalculste my routes, now that I am 'off route'. It's a personal thing, I just prefer to sort the deviation out for myself, rather than have the device do it.

My friends set their devices to promoted, too. They though said 'Yes' when asked if they would like the devices to recalculate. Here things became interesting:

1. Most often their two devices would recalculate no problem. I can only assume that this was because there was only one logical alternative road to take around the diversion.

2. Sometimes their two identical devices calculated different alternatives from each other. We know this as they are in radio contact with each other. This matches in with the slight difference we saw when they both imported otherwise identical routes. In other words, two seemingly identical devices worked slightly differently. Of course it's easy to guess - maybe even correctly - that the reason for the differences in the recalculated routes through the diversion was nothing more than they pushed ther 'Yes' to recalculation at different times, so their position on the map had changed, bringing with it other alternatives.

3. Most often the recalculated routes were perfect. Sometimes though they were absolutely bizarre, routing them sometimes miles away to rejoin the pre-set route, through what was otherwise just a simple (say five miles maximum) deviation from a set route due to roadworks. Had they followed the recalculated route, they would have - literally - been taken all around the houses.

Having in mind that there were no waypoints in any of the routes A to B, we can rule out the obvious possible conclusion that the devices were simply trying to reroute through a 'must do' point we had missed out. The second conclusion is that the two devices were recalculating bizarre alternatives in order to take in a shaping point. But that, even if it were the case, seems odd. I was though reminded that the Grey One once had a bizarre recalculation, going all over the place, too.

Just as an experiment, I tried a prompted recalculation once, in a big town with roads closed. The deviation was easy enough to follow in the town: Turn left, ride about half a mile, turn right, turn left and you'd be back on route, the road closure avoided. It's the way I went, following the temporary road signs, let's say not much over half a mile, a mile at tops, all on simple town streets. The recalculated route my Nav V offered up was, well the word 'bizarre' does not do it justice. It suggested I leave the town, ride miles, pick up a motorway, leave the motrway, re-enter the same town along the road I had come in on and, I assume, had I followed it encounter the same deviation and then would continue on repeating the same until I ran out of petrol, lost the will to live or hurled the expensive device into a river. Why it offered up this strange alternative have no idea, I can only assume that it was to take in a shaping point that was now behind me? Being a bright spark, I ignored the device's instructions to waste two hours of my life and rode to the correct part of my pre-set route, which the device then took up quite happily.

I think I am pretty good at plotting routes, whether in BaseCamp or Mapsource, working Garnin GPS devices and patient enough to learn (or at least try to understand) their occaisional foibles. I am though still no big fan of recalculation, not least as I can't always understand how the device makes the recalculation a reality. In other words, I cannot for the life of me understand why the device, instead of rerouting me very simply through the town (it really was very simple) offered up an alternative that was so crazy it defied even my fervent imagination. I will ask Garmin when I come back, sharing with them the route and the track of the simple track I took, instead of blindly following the devices bonkers alternative offering.

What I haven't mucked about with is the Nav V's 'Devation' button, which allows deviations of up to half a mile. I assume it's some sort of limited recalculation? If that is indeed what it does, then I guess it would have been perfect for the deviation in the town, perhaps?
 
Sat Nav can be the work of the devil

I use them, but mainly on scrolling mode (not doing a route)

I still much prefer a map in my Baglux map case to follow and a pair of eyes to read the road signs
 
Very interesting read that was Wapping, thanks for taking the the and trouble to highlight this.

I think I have had my nav V coming up to a couple of years now and most times when I use it , I learn something new on it. They are superb machine even if they are a little complicated at times
 
Myself and friends had similar experience a couple of weeks ago whilst in Luxembourg and Germany. At least three of us had Nav 5 satnavs with the same latest maps and the same settings. A route created by me and shared to them via the bluetooth feature often showed some changes - even in one case a left turn as opposed to a right turn as we left the hotel!

I don't agree with John Boxer that satnavs are the work of the devil although I do think there is a certain amount of magic involved in their operation - talking to satellites thousands of miles away, is that for real? - and perhaps sometimes they have an element of human thinking resulting in them making different routing decisions depending on their monthly cycles or biorhythms

Wouldn't like to navigate some of the cities or complicated motorway interchanges without one however. I shudder when I think of returning to the days with paper maps ...but we managed
 
Sat Nav can be the work of the devil

I use them, but mainly on scrolling mode (not doing a route)

I still much prefer a map in my Baglux map case to follow and a pair of eyes to read the road signs

Whilst possibly true in part, it does nothing to answer my question, JB.

PS I too can read a road sign, have more than one map (and a handwritten route) and can use my imagination too, me old fruit. It's why I didn't ride miles out of my way in the town and set my recalution to off by default. Others differ.

We followed some very small roads in the wilds of Germany, cross-crossing the countryside, very often with no road signs or with signs to places that were not easy to find even on a detailed ADAC map. My device behaved faultlessly, guiding me hither and yon, leaving me free to hoon the awsome steed without so much as a care in my life.
 
Whilst possibly true in part, it does nothing to answer my question, JB.

PS I too can read a road sign, have more than one map (and a handwritten route) and can use my imagination too, JB. It's why I didn't ride miles out of my way in the town and set my recalution to off by default. Others differ.

We followed some very small roads in the wilds of Germany, cross-crossing the countryside, very often with no road signs or with signs to places that were not easy to find even on a detailed ADAC map. My device behaved faultlessly, guiding me hither and yon, leaving me free to hoon the awsome steed without so much as a care in my life.

I just use the sun to guide me, gets tricky when it is raining though ....although a few will testify that have followed me, that I have my own inbuilt sat nav somehow

I like a combination nowadays - but much prefer a map in an A4 map case
 
A few years ago. Three Garmin SatNavs on three bikes. Two new identical ones and mine which was an ancient 2610. We would come to a cross roads and the respective satnavs several times each chose a different alternative. After a few stops following wild panicked gesticulations from the following riders we decided that we would just go with whomsoever was leading irrespective of what our own SatNav said.

It did fail when we lost sight of the leader on an French roundabout leading to a motorway. He went one way and we another. We only met up some twelve hours later in Switzerland. (Both my colleagues had either lost or ruined their mobile phones before we had even left the UK!!!)
 
I'd guess that your devices were all set to 'fastest', so when a route is recalculated the sat nav has to decide on which roads are likely to be the fastest (starts looking first at arterial routes?) and may then have to calculate the fastest route at branches (in other words decision points) based on the distance and the speed limit data coded in the map set. In order to do all this quickly, and whilst the actual position of the satnav is changing (because no one ever stops :D) there must be a fair amount of approximation in the calculations. So ending up with somewhat different routes on near-identical devices isn't too surprising.

All conjecture of course but it does highlight the importance of a couple of basic rules:

1. the leader is always right, even if everyone else's satnav shows different.

2. always, always check the route calculated by the device, including diversions calculated on the fly. One of the big benefits if the scroll wheel on the LC GS is that it allows rapid zooming in and out which is invaluable when looking for strange quirks in routes. For instance, I've found that several Garmin models (550, 220, Nav V) will sometimes take the route off at a motorway junction and then straight back on again. Weird, and very irritating but easily avoided by zooming out a little.
 
We had a similar experience on our recent trip. I authored all the routes in Basecamp, using a combination of waypoint and shaping points. I needed waypoints as in some circumstances no amount of route dragging took me the way I wanted to go - for example I had to put about 4 waypoints in to navigate the Transfagarasan, which basecamp seemed to hate.

Anyway, we had a Garmin 595 on my bike, a Garmin 590 on Mrs Berins bike and a spare 590. The 595 and spare 590 had 2017.1 installed, and Mrs Berin had 2016.4? (the previous version, anyway) I also had Bulmaps and adriaroute installed.

I sent the routes to the devices, but we tend to only import the routes we'll need for the next day, and set off. In Western Europe the two devices in use pretty much agreed, but once we got past Czech differences started to appear. Also, the 595 has irritating warnings which continually flash up about animal crossings and schools and such, but they stopped after Slovakia. Anyway, the animal crossing warning is useless - a moose ran out in front of me and there was absolutely no warning of it on the sat nav! The route guidance that usually appears on the top bar also stopped.

After a while, it got silly - I looked at the routes in the morning and they looked more or less the same, but then Mrs Berin's device would have a fit, spend 30 minutes recalculating and then say she has 5683 miles to go that day! Stopping and restarting the route usually fixed it, but ecru weird. I reasoned it may be due to different maps, so we swapped for the spare device with the same maps- and the same thing happened. We had comms so we compared routes as we went, using mine as the master, and there were often differences.

A couple of times my 595 switched off by itself, and a couple of times wouldn't switch back on after it had switched off when the bike was switched off -reset by disconnecting the battery fixed it.

As we got back to civilisation we started getting stuff like route guidance and lane guidance back (but not at the same time on each device) and the routes started disagreeing less. I also got the naggy school crossing warnings back.

A few times when we had to replan we just put in a destination and although usually the route was the same (we set route calculation, avoidances etc to be the same) it wasn't always.

But as Wapping said, a trip like this would have been very difficult without them - we went through 15 countries and to take detailed maps for all would have filled all our luggage space, and we switched between dirt roads I'd found on Google Earth and motorway blasts to cover ground without having to stop every 25 miles to reposition the map!


Sent from my iPad using Tapatalk
 
What I haven't mucked about with is the Nav V's 'Devation' button, which allows deviations of up to half a mile. I assume it's some sort of limited recalculation? If that is indeed what it does, then I guess it would have been perfect for the deviation in the town, perhaps?

While away recently in the Loire we came to a route barree I decided to try the deviation function to see how effective/useful it was.

The 590 calculated some random route over horrible minor roads for about 15 km out of my way, my GPS is set to faster time so have no idea why it chose minor country roads.

Fortunately I had set a track of my route which I always do, all I did was work my way past the deviation by zooming out, then get back on to the track.

Wont be using it again!!
 
While away recently in the Loire we came to a route barree I decided to try the deviation function to see how effective/useful it was.

The 590 calculated some random route over horrible minor roads for about 15 km out of my way, my GPS is set to faster time so have no idea why it chose minor country roads.

Fortunately I had set a track of my route which I always do, all I did was work my way past the deviation by zooming out, then get back on to the track.

Wont be using it again!!
This has happened to me on my 660, taking me over roads where average speed was down to about 20mph on my fully laden 1150.

After a bit of faffing about I found that changing the 'vehicle' from bike to car improved things no end, it would seem that whoever programmed the thing assumed that bikes always want to travel via (very) tiny back lanes, this was in France.

I now always have my 660 set to 'Car'... though I nearly always use custom routes, but it is handy when encountering road closures.
 
I just use the sun to guide me, gets tricky when it is raining though ....although a few will testify that have followed me, that I have my own inbuilt sat nav somehow

I like a combination nowadays - but much prefer a map in an A4 map case

The sun, which has shone solidly all the way around Germany for the past three weeks, is a great guide. In your face in the early morning, you are going east. In your face in the evening, you are going west. Over your shoulder, bleaching out your satnav at noon, you are going north. Straight into your dark visor at noon, you are going south.

Rivers, sea, lakes, mountains and their position relative to you; great clues, too. But only really helpful if you know where they are meant to be.... Which is where a map always helps.
 
I'd guess that your devices were all set to 'fastest', so when a route is recalculated the sat nav has to decide on which roads are likely to be the fastest (starts looking first at arterial routes?) and may then have to calculate the fastest route at branches (in other words decision points) based on the distance and the speed limit data coded in the map set. In order to do all this quickly, and whilst the actual position of the satnav is changing (because no one ever stops :D) there must be a fair amount of approximation in the calculations. So ending up with somewhat different routes on near-identical devices isn't too surprising.

All conjecture of course but it does highlight the importance of a couple of basic rules:

1. the leader is always right, even if everyone else's satnav shows different.

2. always, always check the route calculated by the device, including diversions calculated on the fly. One of the big benefits if the scroll wheel on the LC GS is that it allows rapid zooming in and out which is invaluable when looking for strange quirks in routes. For instance, I've found that several Garmin models (550, 220, Nav V) will sometimes take the route off at a motorway junction and then straight back on again. Weird, and very irritating but easily avoided by zooming out a little.

Spot on, not least about the leader being always right.

I thought about the fastest time thing. The route through the town (with or without the diversion) was say two miles. The speed limit is, at worse, 20 MPH. That should equate to six minutes. There is no way on this earth that (even I) could have hooned the awsome steed around the recalculated route in six minutes or under.
 
could it be that the unit is trying to take you to the next closest point of reference, and that may have been the start of the route or somewhere that was a close point to it, and then routing you back down to your end waypoint using your own preferences, that may or may not have taken your own route.
 
Before leaving uk we had done several local to home runs on pre planned routes. 590 on mine and 660 on Julie's bike. All boxes unchecked on satnav and basecamp. Found that if I made routes in mapping using the ones downloaded for 590 would alter when loaded onto 660. If mapping for 660 used and saved to each unit very little change. Any changes and Julie following 660 would recalculate and follow my route.
When on our euro tour all routes pre done as above for each day,loaded to units as above. When coming across the many diversions,using the detour didn't always work, my 590 wanted to reroute fastest to end destination. We then always followed diversion signs,turning route off and back on once on track. We had maps for each region too so no problems. I did redo a route to go from Hann Münden through the Harz and on to Paderborn using car setting and had less issues. The 660 had less hissy fits and on one occasion in Black Forest I just followed the Mrs,most relaxing.
I am wondering if there is less road coding when on minor roads whilst abroad and in cities who knows.
 
could it be that the unit is trying to take you to the next closest point of reference, and that may have been the start of the route or somewhere that was a close point to it, and then routing you back down to your end waypoint using your own preferences, that may or may not have taken your own route.

That's a pretty good guess. Quite why it wouldn't re-route me to the next known shaping point which was about a mile away in the town and from there to the rest of the route, I am not sure. I have a nagging feeling that the modern devices - like the Nav V and its Garmin equivalent - like to have a few waypoints built into custom routes, as opposed to just lots of shaping points. With my A to B route there was only two waypoints, the departure hotel and the arrival hotel, separated only by my custom made route consisting entirely of shaping points.

The odd thing then is thst my two friends also made recalculations; sometimes their devices made them perfectly, sometimes they didn't.
 
Is it also possible that the shaping points that are created in basecamp are not uniform or regular but more just happen where you get hold of the route and drag it.
The software then interprets that in some way differently.

If firm way points had been used ( i never use them myself) the re route may have been more logical
 
. . . although I do think there is a certain amount of magic involved in their operation - talking to satellites thousands of miles away, is that for real? - and perhaps sometimes they have an element of human thinking resulting in them making different routing decisions depending on their monthly cycles or biorhythms

Ill go along with that. Last month I had 10 days in Germany through IJmuiden and my Zumo 660 generally behaved well, but had a few complete hissy fits with 2,000 km to unknown via point. And a couple of bizarre choices.

Like others, I've learned to be cautious with via points, be wary of route recalculation (PC to Device and planned routes) and planning routes on the device.

All in all they are challenge, but also wouldn't be without them. Impossible to contemplate riding solo with only a large map for long journeys.
 


Back
Top Bottom